Adapting to the fast pace of a live multiplayer VR game with PIGIAMA KASAMA

Founded in 2022, PIGIAMA KASAMA is a game production company that creates games for its own IPs, such as Kasama: The Awakening, and offers game development services to other brands.

Their newest title, PICKABOOM, is a next-generation social VR game built around exploration, chaotic interactions, and real-time multiplayer cooperation. Players descend into unpredictable cave systems, survive absurd natural disasters, and create their own moments of comedy, tension, and triumph.

We spoke with cofounders, Matteo Fanchin and Paride Stella, to discuss how they approached networking instability, reconnection failures, performance drops, and quick content turnaround times – all while building their first live multiplayer VR game.

What was the main goal with PICKABOOM?

Paride Stella: Our main goal was to create a high-energy social VR game that feels alive, unpredictable, and endlessly replayable – a world where players team up, survive together, and naturally generate the kinds of moments people want to share.

As a live game, PICKABOOM has fast-moving demands: we release updates weekly, constantly fix issues surfaced by the community, and add new content while keeping performance solid. This required us to design systems that scale, synchronize physics across multiple players, and stay performant on standalone headsets.

We wanted it to feel like chaotic co-op survival mixed with creator culture, where the environment, disasters, and physics interactions constantly surprise players and push them to collaborate.

What was the biggest technical challenge with developing PICKABOOM?

Matteo Fanchin: PICKABOOM is our studio’s first multiplayer title, and it pushed us to confront networking instability, reconnection failures, performance drops during chaotic physics events, and the pressure of shipping weekly updates while continuously improving gameplay quality.

What steps did the team take to overcome it?

MF: We broke the problem into stages – building a live content pipeline and implementing continuous profiling. For the live content pipeline, we modularized tools and enemies, allowed new content to be dropped in quickly, ensured updates wouldn’t break previous builds, and adjusted the weekly schedule when necessary.

For continuous profiling, we ran regular performance passes, tested multiplayer load with real community players, and prioritized VR-specific optimization for memory usage, garbage collection, and GPU spikes. This approach allowed us to ship stable weekly updates without slowing development.

What new features and updates in Unity 6 helped overcome the technical challenges the team faced?

PS: While scaling the game, we streamlined iteration time and improved performance with the following features and updates:

Faster build times: A live game requires frequent builds. Given the multiplayer nature – up to 10 players simultaneously – it was crucial to test with both internal and external teams. Faster build times saved a significant amount of development time.

Universal Render Pipeline (URP) performance improvements: Developing for VR is always challenging due to hardware limitations. Improvements to URP made development easier, as the performance gains were tangible. We tuned lighting, shadows, and materials for clarity and readability in chaotic environments.

Shader Graph upgrades: We used Shader Graph extensively to improve asset creation speed, procedural effects, character customization, and optimization.

We developed custom map shaders to procedurally paint complex materials with minimal performance cost. The shaders can generate multiple visual biomes. They create dramatically different looks while using very few resources and maintaining a low impact on GPU and memory.

Shader Graph was also instrumental in building our avatar material system. Through a custom pipeline, a single material can be shared across all characters without relying on textures, while still allowing players to customize nearly every element of their avatar – down to details like nail color. This approach ensures maximum performance even when large numbers of characters and accessories are displayed simultaneously.

Unity Physics performance improvements: Unity Physics powers all interactions, including throwing, grabbing, enemy reactions, and disasters. We optimized Rigidbody settings to remain deterministic enough for multiplayer.

Unity Audio system improvements: The Unity Audio system was used for positional cues, player voices, enemies, and explosions, helping players coordinate and communicate in real time. These improvements enabled clear spatial sound cues, even in chaotic moments.

Addressables stability: We used Addressables to support live updates of characters, objects, and accessories, ensuring maximum flexibility and enabling long-term scalability for the project.

How did the team implement GPU-optimized effects?

PS: We implemented GPU-optimized effects by first using GPU instancing for vegetation and gems. This played a major role in reducing the performance impact of VFX and particles. We then optimized each shader to keep it as lightweight as possible, containing only the logic strictly required for its purpose.

We also optimized geometry to keep polygon counts and draw calls as low as possible. Our intricate cave system is extremely lightweight, allowing smooth performance on standalone VR hardware. Large sections of the PICKABOOM map can be prototyped, tested, and deployed in under two working days thanks to our custom-developed map editor.

Lastly, we reduced lighting costs. The game’s visual style was designed to minimize the use of real-time lighting almost entirely. Real-time lights are used only when absolutely necessary, such as for specific VFX moments or the player’s torch. Our custom map editor includes an adaptive light probe system built specifically for cave environments, allowing us to achieve stable, believable lighting at a fraction of the performance cost.

We benchmarked each part of this approach for 60+ fps on Quest 3 and Quest 3S.

What are your content stream goals?

MF: We release new updates every week, including new items, new mini-games, stability improvements, and player-requested features. We plan two to three weeks ahead but maintain flexibility based on player feedback and crash reports. We’re able to turn around content very quickly by using modular content, rapid prototyping approaches, and Addressables.

What performance-related issues did the team encounter?

PS: We faced physics spikes, Rigidbody explosions, network jitter, and GPU load issues during disaster gameplay. We addressed these challenges by simplifying the physics layer, optimizing colliders, using GPU instancing and lower-cost shaders, improving synchronization frequency, and removing unnecessary object updates.

The Unity Profiler and the Frame Debugger were very helpful. We also relied on the Deep Profiler to pinpoint garbage collection issues, as well as Android Logcat for runtime diagnostics.

What milestones are you most proud of?

MF: First and foremost, we’re happy to have achieved a stable 60+ fps on Quest 3 and Quest 3S. We’re also proud to support hundreds of networked physics objects without any performance loss and to maintain zero “blocker” bugs during weekly updates. Lastly, it’s great to have a camera system that captures high-quality footage without frame drops.

What advice would you give a developer looking to scale their VR game?

PS: My advice would be to prototype quickly and refine later. It’s also important to keep multiplayer architecture simple before scaling and to profile constantly.

MF: It’s essential to build content in modular pieces and to prioritize frame stability over visuals. I’d also recommend testing with real players weekly and developing tools early.

To read more about projects made with Unity, visit the Resources page.